I would think that fixed wing wouldn't be very practical for airsoft because of the difficulty of picking out players in camo with a fairly fast pass capturing limited resolution. A platform which can loiter in an area would force players to freeze as long as an area was being watched. A fixed wing platform would not provide a very long look time.
Another concern I would have with fixed wing UAVs would be their higher flight speeds. A fast moving plane crashing down would come down pretty fast which I think could be a safety hazard. I've had a few R/C plane crashes that left some major divots in the ground. That being said, large 0.60cuin R/C helicopters can be bloody scary too. I hated it when those friggin things would be tootling around at head height across the flightline whirling their massive blades. It seems to me that a lightly built electric quad rotor UAV wouldn't present nearly the same kind of collision damage potential as a fixed wing plane blazing around.
That being said there are some cool benefits to fixed wing UAVs. They're easy to make strike tolerant. I've flown foam winged planes with lots of superficial poke holes in the wings (encounters with lighter tree branches) just fine. Stout foam wings can take a lot of damage without being compromised and they're cheap to replace. Major structural elements like spars are strong and would probably be difficult to damage with an AEG. It would be very hard to get a critical strike on a plane if you put the receiver in a plastic box and string the antenna through a plastic tube so it doesn't get snipped. I guess the only thing you could hit would be a control surface hinge, but good luck doing that. All you'd have to do is put a thin sheet of plastic on the underside guarding the hinge to assure that control surfaces continue to work under fire.
For safety I would limit fixed wing UAV platforms to stable high wing or high dihedral models with long chord foam wings (lots of material to chew out before it's compromised). Limit displacement to maybe 0.25cuin to keep the engine block light and maybe even require pusher arrangements so if someone gets nailed by the plane they at least don't get hit by a spinner and hunk of engine. The nose of the plane should not be sharp (say minimum 1.5" radius semispherical nose) and made of a soft nerfy foam rubber. We would not want exciting dogfighting models out there which might facilitate or encourage reckless flying at high speeds. If we can keep the speeds down the remaining risk I see is goggle knockoff. A slow plane to the face can dislodge goggles which is a hazard if the victim is also receiving fire.
If you think r/c planes are easy to shoot down, check this shit out:
YouTube- RC Plane vs AR15 Machine Gun
Technologically speaking a fixed wing UAV would be easy to implement, but we should assess the safety issues carefully. It might even be fun to have something that we could intentionally shoot at. DonP did some very interesting work with electronic hit detection on airsoft targets. He might be able to detect pellet strikes to a RC plane which lights off a smoke charge. We can also do some milsim fun like do up an olive drab model Piper Cub as an L-4 artillery spotting plane. In a particular scenario if you can't shoot down the spotting plane in a certain amount of time some offshore or out of ranged artillery turns your base into aggregate.
For outright spotting utility, I think a rotary wing UAV would work best. I'm thinking of buying this Walkera quadcopter:
http://www.walkera.com/en1/particular.jsp?pn=UF0+4%23
Seems pretty cheap and it's apparently camera ready. Unfortunately I have a feeling that it's flight time isn't anything spectacular (I'm guessing 7min). A combustion engine can easily achieve much longer, especially on a fixed wing platform. Another benefit of an IC engine is that refuelling is pretty damn fast compared to battery charging. However, I think being able watch an area without the entire frame moving would facilitate observing camouflaged players. I'm not meaning to pick out players in the forest, but a guy in green in tall grass may be hard to pick out with a fast rc plane pass. Also it's hard to get a sense of location when you're getting multiple strafing passes from different directions.
Maybe we need to think out of the box a bit harder. Perhaps we could put a kickass lipo on a helium filled derigible which powers some pretty decent brushless motors. If we could beat a not bad crosswind with some good motors we might be able to achieve some good uptimes with a pretty stable platform. It would be bad to shoot at a derigible, but we could maybe stomp on the rifle on whomever decides to take potshots at a cool item that a game org charitably provides.